Transcript: 'The First Step Toward Building an AI-Ready Accounting Firm'

‘Accounting Tech Talks’ with Credfino’s Dheeraj Pandey

The transcript of Accounting Tech Talks with TaxFlow founder Ben Vickery, hosted by Dheeraj Pandey, is below. Watch on YouTube or listen on Apple Podcasts.

Timestamps

  1. Introduction to Ben Vickery and TaxFlow: 00:00:00
  2. How a spreadsheet problem led to TaxFlow: 00:00:37
  3. Do smaller firms need full practice management software?: 00:02:06
  4. How TaxFlow supports smaller firms and internal tax teams: 00:02:57
  5. Why accountants should start thinking like engineers: 00:07:06
  6. The reality of AI adoption in accounting firms: 00:10:04
  7. Why accounting firms aren’t ready for AI agents: 00:11:35
  8. How accountants can safely experiment with vibe coding: 00:13:57
  9. When to involve a fractional software engineer: 00:15:21
  10. How AI can help accountants become better advisors: 00:16:28
  11. What the accounting firm of 2030 might look like: 00:19:39
  12. Why accounting firms may become more like software companies: 00:20:07
  13. The growing role of fractional engineers in accounting firms: 00:21:06
  14. New opportunities for engineers in professional services: 00:22:55
  15. Closing thoughts: 00:24:03

Transcript

(00:00:00)

Dheeraj Pandey

Hi, welcome to Accounting Tech Talks.

It’s a podcast where we talk about the latest developments in accounting and technology, particularly in the U.S. accounting industry.

Today, I have Ben with me, who’s the founder of a software product called TaxFlow.

We’d love to understand some of the challenges his product solves and hear a little bit about Ben’s journey.

Ben, welcome to the podcast.

Ben Vickery

Thank you.

Dheeraj Pandey

Great to have you here.

I’d love to understand a little more about what led you to build TaxFlow.

(00:00:37)

Ben Vickery

Yeah, absolutely.

My journey started when I was working at a local CPA firm here in Colorado.

We already had a lot of our tech stack in place. We had a good client portal and good communication with our clients.

But what we were struggling with was our workflow.

We were still using a spreadsheet to track where each return was in the process.

What’s the status of this return? Are we waiting on information? Who’s working on it?

Those kinds of things.

So I started looking for a solution, and what I found was that my only real option seemed to be a full practice management system at the time.

I was also getting into programming, so I decided to build a really simple, streamlined workflow system for our firm.

That’s actually how TaxFlow was born.

We didn’t want to replace everything.

Our clients and staff were already trained on the tech stack we had in place.

We just wanted to solve the problem of tracking everything through a spreadsheet, especially in the age of cloud accounting.

I ended up building it for ourselves, and it was really useful.

From there, I thought maybe other accounting firms could use a simple workflow tool instead of purchasing a full practice management system and changing their entire tech stack.

So I started reaching out to more accountants and showing it to them.

It snowballed from there and eventually turned into the TaxFlow we have today.

(00:02:06)

Dheeraj Pandey

Great. That’s good to know.

I love the fact that you focused on simplicity.

But one of the challenges I see right now is that practice management systems are becoming broader.

They’re adding features such as smart intake and even tax return delivery, which weren’t necessarily part of their original offerings.

So, just an intuitive question.

Was this something you encountered when you started showing TaxFlow to different firms?

Or is your audience really solo practitioners and very small firm owners who simply need something to move beyond Excel?

(00:02:57)

Ben Vickery

Yeah, that’s a great question.

We generally serve firms with around one to 15 people.

So we’re definitely focused on the smaller side of the accounting firm market.

But what I’ve found is that a lot of it comes down to the personality of the firm owner.

Every firm is made up of different people, which makes every firm unique.

That carries over into their technology decisions and how they like to work.

Another important factor is that many firms already have technologies in place.

Changing those systems becomes a procedural issue.

They start thinking, “Are we really going to remove all these different features and implement an entirely new tech stack?”

Then everyone needs to be retrained, including the clients.

Even though practice management platforms keep adding more features, some firms are happy with their existing tools and their current client base.

They just want to improve certain operational processes.

They don’t want to jeopardize their workflow or completely change how their firm operates.

They want to get out of spreadsheets without changing their entire firm and its processes.

So that’s something we definitely see.

There are still plenty of firms saying, “We really just need a better workflow system because we’re tired of using spreadsheets.”

(00:04:23)

Dheeraj Pandey

Understood.

So you’re focusing on firms that may feel overwhelmed by technology.

Perhaps smaller or solo firm owners who find technology intimidating, or firms that don’t necessarily want to purchase a large practice management platform.

They might have a small team that wouldn’t fully use such a system, or they simply don’t have the time to understand and implement a more complex platform.

Would that be a fair understanding?

Ben Vickery

Yeah, absolutely.

That’s one side of it.

The other segment we discovered after launching TaxFlow was internal tax teams.

They may not need many of the features included in an accounting practice management system.

So we serve both groups.

On one side, there are smaller firms that want something simple, perhaps because they’re less tech-savvy or want a solution that works alongside their existing products.

On the other side, there are internal tax teams that don’t necessarily need all those additional accounting features.

Those are the two segments we focus on.

Dheeraj Pandey

Understood.

So how does the tool actually work?

Now that we understand who uses the product, tell us a little about its functionality.

(00:05:46)

Ben Vickery

Yeah, absolutely.

It’s a really simple workflow system.

One thing we offer is support for multiple entities.

You can have an account level and then individual client or entity records underneath it.

Let’s say you’re working with a client and preparing a Form 1040, but you’re also handling a trust, estate, or company return.

You can track all of that under the same account.

On top of that, we have workflows where you can preassign users.

For example, you can specify who’s responsible for preparation and who’s responsible for review.

You can set up those workflows in advance and then check off tasks as they’re completed.

When you’re the next person responsible for a task, you receive a notification.

It keeps everyone on the same page.

You know when it’s your turn to complete something.

Once a task is checked off, there’s an audit history.

You can also communicate with other people inside your firm through notes.

It’s a very simple way to keep everyone informed while still having useful software features such as notifications and internal communication.

The main feature set comes down to a few questions:

Who’s working on what?

What’s the current status?

When is it my turn to complete my work?

Do I need to ask someone else on the team for additional information?

That’s the general workflow and project management problem we’re trying to solve.

(00:07:06)

Dheeraj Pandey

Okay.

It looks like you’ve found a sweet spot and a good balance between functionality, ease of use, and internal team communication.

That’s great.

I’d like to talk a little more about what you’re seeing in the accounting industry at large.

We’ve been reading about your perspective that accountants should start thinking like engineers.

On one hand, we see that accountants sometimes find technology difficult or overwhelming to adopt.

So what does it take for accountants to move from that mindset toward thinking a little more like engineers?

What’s your take on that?

(00:07:55)

Ben Vickery

Yeah, absolutely. That’s a great question.

When I say accountants need to start thinking like engineers, I’m really talking about two things.

One is developing an iterative mindset.

The other is developing a testing mindset around that iteration.

If you look at software development from a software engineer’s perspective, we rarely release a completely finished feature on the first attempt.

A lot of the time, we’re talking about a V1, or version one.

Then there’s a V2, or version two.

The whole idea is to create that first version, get feedback, test how it performs, and then iterate.

That process helps you develop version two, version three, and eventually something that works really well.

The same mindset applies to accounting, particularly in the AI era.

I often see accountants using AI for the first time.

They might set up a skill, create a scheduled task, or try something similar.

Then they see the first output, and it’s not quite what they were looking for.

So they conclude, “AI is just hype. This isn’t going to work. I don’t understand why everyone is so excited about this.”

But my point is that you need to treat that first output as version one.

You need to see what the system produces.

If you don’t like the output, adjust the prompt you’re giving the AI.

When you get to version two, you’re hopefully a little closer to what you’re looking for.

That’s where the testing mindset comes in.

You’re testing the outputs internally to see whether they’re meeting your expectations.

By the time you reach versions three, four, or five, you may have a really useful system that provides substantial value.

But you have to iterate to get there.

That’s how we approach software engineering.

We build the first version, gather feedback, test it, improve it, and continue iterating until we have a useful final product.

So when I say accountants should think like engineers, I’m talking about adopting that process.

Build a first version, test it, iterate, and keep improving it until it becomes something that genuinely helps your firm.

(00:10:04)

Dheeraj Pandey

Are you seeing a lot of accounting firm owners experimenting with AI products now?

For example, are they using Claude Code, Claude Cowork, or similar tools to build something meaningful for their firms?

Are you seeing that happen frequently?

Ben Vickery

That’s a really good question.

I think it’s still pretty early.

People are starting to get into it, and adoption is growing.

But when I talk to people signing up for TaxFlow, I still come across firms that are physically moving paper files around their offices.

There’s such a wide spectrum in the accounting industry.

If you’re spending a lot of time on LinkedIn, you might assume everyone is using these AI products.

But when you’re actually talking to accounting firms, you realize there’s still a lot of work to do to bring everyone up to that level.

We need to help firms reach the point where they can start building with AI.

The market is really all over the place.

There are people at every stage.

But I would encourage everyone to start trying it.

Start experimenting and see what you can do with it.

One of the interesting things about AI is that the barrier to getting started is relatively low.

You can type into it or talk to it and ask for help.

You can even ask AI how it might be useful for your work.

It’s actually pretty good at helping you think through potential use cases.

So I’d encourage people to start using it, ask questions, and explore what it can do.

It’s really exciting technology.

(00:11:35)

Dheeraj Pandey

You also mentioned in one of your posts that accounting firms aren’t ready for AI agents yet.

That’s an interesting take.

Why would you say that?

Ben Vickery

Yeah, that’s a great point.

Think about what I was saying a moment ago.

I still talk to firms that are moving paper files around the office.

We need to get those firms into the cloud before they can start using AI across their operations.

But what I really mean is that there’s a prerequisite to getting the most out of AI.

And that is understanding your internal processes.

Inside an accounting firm, it’s entirely possible that three different people follow the same procedure in slightly different ways.

What we’re trying to do is get those processes out of people’s heads and into a system.

That way, when you train an AI tool, it understands what you want it to do.

This is where the idea of treating AI like a junior staff member comes from.

You wouldn’t give a junior employee extremely vague instructions and expect them to complete a task perfectly.

You’d give them the client file and explain the procedure your firm normally follows.

You need to approach AI with that same mindset.

Start documenting your procedures.

There are several ways to do it.

It might sound overwhelming to say, “We need to document every procedure we perform in our firm.”

But AI is also making that documentation process easier.

So step one is getting those processes out of your head and into a format that an AI system can follow.

That’s how you enable AI to work according to the way your firm actually operates.

(00:13:27)

Dheeraj Pandey

That’s a very interesting point.

I think small firm owners, especially those running founder-led firms, struggle with this quite a bit.

A lot of the processes and context live inside the founder’s head.

That’s why, even if someone wants to use AI or create something meaningful, it can be difficult to automate their practice.

You’ve also spoken about vibe coding.

I’ve seen people on both sides of that discussion.

Some people love it, while others say it isn’t everyone’s cup of tea.

What would be a good way for accountants who are a little more tech-savvy to experiment with vibe coding?

Are there particular platforms or approaches they can use?

(00:14:25)

Ben Vickery

Yeah, that’s a great question.

There are several ways to approach vibe coding.

I think it’s worth experimenting with.

One useful tip is that AI is really good at generating synthetic client data.

That’s a great way to set yourself up to test something through vibe coding.

You probably don’t want to start building applications with real client data if you don’t understand security practices or how to design a secure application architecture.

But you can generate fake client data with AI and build your prototype around that.

My approach is to prove the value for your firm first.

If you can see that connecting actual client data to a particular application would create meaningful value, then you have something worth exploring further.

(00:15:21)

Ben Vickery

At that point, it might be worth bringing in a fractional engineer.

Maybe you pay someone $500 to $1,000 to audit the application you’ve created through vibe coding.

The idea is to make sure it’s secure before you start processing actual client information through it.

But I’d also say that if you’re interested in vibe coding, you don’t necessarily have to start by building something for your firm.

You can build something you simply find interesting.

That’s a great way to learn.

For example, when I was learning Claude Code, I built an accounting memes website.

It used AI to generate accounting memes.

I didn’t want to experiment on a project people were already relying on.

I wanted to learn how to use the technology.

So I built something fun that I could put out there for people to use.

You could do something similar.

Build a personal project and get comfortable with the tools.

Then, once you understand how they work, you can start applying those skills to your actual firm and developing internal tools.

(00:16:28)

Dheeraj Pandey

Understood.

That’s a great perspective on vibe coding and how accountants can experiment with AI and software development.

You’ve also spoken about what the accounting firm of the future might look like.

How do you think AI can help accountants become better at advisory services?

Everybody seems to acknowledge that advisory is an important part of the profession’s future.

How do you see accountants using AI to deliver better advisory services?

(00:17:08)

Ben Vickery

Yeah, that’s a great question.

To your point, that’s what everyone is talking about.

We’re automating a lot of back-office work.

AI is beginning to handle more technical tasks.

So the question becomes: how do we make the human professional better in that environment?

How do we improve advisory skills while more of the back-office work is being automated?

It’s really interesting.

I’m working on a product called Navi, which Jody Padar founded.

What we’re focused on is helping people improve their advisory and communication skills in a world where more technical work is being automated.

One of the reasons I’m excited about this product is that something like this wasn’t really possible five or six years ago.

But now we can record calls and generate transcripts.

We can take those transcripts and analyze them through software.

AI can help identify moments where you could have asked a better question or gone deeper into a client’s response.

It might tell you, “The client asked a question like this, and you should have explored it further.”

Or it might identify that you’re being too technical.

How can you bridge that communication gap so your client better understands what you’re trying to explain?

You can also identify potential new opportunities, such as additional services or possible upsells.

If you think about it, the professionals who are better communicators and better advisors are often the ones who can charge a premium for their services.

And those abilities are becoming more important as technology automates more back-office work.

So how do you become a better communicator?

How do you show up for your clients in a way that genuinely helps them understand what’s happening in their business?

That’s what we’re trying to address with Navi.

You can take your call transcripts and get useful insights into how you’re communicating.

That helps you improve as an advisor while you’re also improving your back-office processes with AI.

The two go hand in hand.

Technology can help improve both operational efficiency and advisory skills.

(00:19:24)

Dheeraj Pandey

Great.

I think it’s good that you’re solving that problem as well.

It’s much-needed technology because firms do need to become better at advising their clients.

Using technology to improve how they serve those clients is an important part of the future.

Let me ask you another question.

If you were looking at a modern accounting firm in 2030, what kinds of roles would you expect to see?

What would the team look like?

What roles would people have in a modern accounting firm operating in that environment?

(00:20:07)

Ben Vickery

Yeah, that’s a great question.

Taking a slightly higher-level view, I can see accounting firms becoming almost like their own software shops.

They’re becoming more technical as they build AI skills and start working with tools such as Claude Cowork and Codex.

A lot of that work is actually quite technical.

I’ve been teaching some Claude Cowork courses for accountants.

What I’ve found is that accountants can definitely learn these tools, but a lot of software engineering concepts are now appearing inside the interfaces they’re working with.

So I see an opportunity here.

Accounting firms are becoming more technical.

At the same time, we’re seeing more software engineers become available in the job market.

Large technology companies are increasingly looking at how AI affects the number of engineers they need.

That creates an interesting opportunity for professional services firms.

(00:21:06)

Ben Vickery

Why wouldn’t an accounting firm consider having a fractional engineer on its team?

That person could help build out Claude Cowork processes and AI workflows.

They could think through how the different skills are structured and when those skills should be triggered.

They could also help manage those systems as they grow.

Going back to the iterative mindset we discussed earlier, you need to think about how you’re maintaining these tools over time.

For example, how do you handle version control?

When you update an AI skill, how do you keep track of those changes?

How do you maintain a record of previous versions?

Those are software engineering concepts.

An engineer could help firms manage those processes more effectively.

There’s another potential use case.

Let’s say you want to vibe code an internal application.

It would be valuable to have an engineer who could take that concept and structure it properly.

They could design the architecture and make sure the application is secure enough to use with actual client information.

So I see two areas where engineers could help accounting firms.

The first is developing and maintaining AI systems, including the kinds of processes you might build through Claude Cowork.

The second is helping firms develop their own internal applications through vibe coding and other development tools.

I don’t necessarily know whether those will become separate roles.

You might have someone focused on AI automation and someone else focused on application development.

But with AI, those roles are starting to converge for software engineers as well.

(00:22:55)

Ben Vickery

I can imagine a future where professional services firms are just as technical as the software companies we’ve been working with for the last decade.

I think that’s a really promising and exciting direction.

I’ve already seen a few firms hire their first dedicated AI professional.

These people are helping set up AI systems and build internal tools.

My expectation is that we’ll continue seeing more of that.

It’s beneficial for accounting firms because they gain access to technical expertise.

And it’s also beneficial for engineers who enjoy this kind of work and are looking for new opportunities.

I think there’s a really interesting convergence happening.

Professional services firms are becoming more technical, and engineers are finding opportunities within those firms.

I expect that to become more normal over the next five to ten years.

Some firms are already taking that step and seeing the benefits of having an engineer on their team.

(00:23:26)

Dheeraj Pandey

That’s a very detailed and relevant point.

I completely agree with you.

I think we’ll definitely see more AI implementation-related roles in accounting firms.

And not just accounting firms, but service-based firms in general.

For people building their careers, it’s also a very interesting opportunity.

Someone working in the software industry could choose to go deeper into the accounting world and become…

[Dheeraj’s sentence is interrupted in the supplied transcript before he completes the thought.]

(00:24:03)

Dheeraj Pandey

Awesome.

Ben Vickery

Yeah, absolutely.

Dheeraj Pandey

Sorry, I think I lost the connection for a second.

Ben Vickery

No worries.

Dheeraj Pandey

But yeah, it was great having you.

We had an awesome discussion.

Thank you for being on the podcast.

Ben Vickery

Absolutely.

I appreciate you having me.

Dheeraj Pandey

Great. I’ll stop the recording now.

You may also like

© 2026 Credfino.com. All Rights Reserved.